Artificial Intelligence Work and Market Power on Online Labour Markets

Abs­trakt wystą­pie­nia:
„We inve­sti­ga­te three alter­na­ti­ve but com­ple­men­ta­ry indi­ca­tors of mar­ket power on one of the lar­gest onli­ne labo­ur mar­kets (OLMs) in Euro­pe: (1) the ela­sti­ci­ty of labo­ur demand, (2) the ela­sti­ci­ty of labo­ur sup­ply, and (3) the con­cen­tra­tion of mar­ket sha­res on the OLM under stu­dy. We explo­re how the­se indi­ca­tors rela­te to an exo­ge­no­us chan­ge in plat­form poli­cy. In the mid­dle of the obse­rva­tion period, the plat­form made it man­da­to­ry for employ­ers to signal the rates they were wil­ling to pay as given by the level of expe­rien­ce requ­ired to per­form a pro­ject, i.e., entry, inter­me­dia­te or expert level. We find a posi­ti­ve labo­ur sup­ply ela­sti­ci­ty ran­ging betwe­en 0.06 and 0.15, which is higher for expert-level pro­jects. We also find that the labo­ur demand ela­sti­ci­ty incre­ased whi­le the labo­ur sup­ply ela­sti­ci­ty decre­ased after the poli­cy chan­ge. Based on this, we argue that mar­ket-desi­gning plat­form pro­vi­ders can influ­en­ce the labo­ur demand and sup­ply ela­sti­ci­ties on OLMs with cor­re­spon­ding recru­it­ment poli­cies. We also explo­re the demand for and sup­ply of AI-rela­ted labo­ur on the OLM under stu­dy. We pro­vi­de evi­den­ce for a signi­fi­can­tly higher demand for AI-rela­ted labo­ur (ran­ging from +1.3% to +4.1%) and a signi­fi­can­tly lower sup­ply of AI-rela­ted labo­ur (ran­ging from ‑6.8% to ‑1.6%) than for other types of labo­ur. We also find that wor­kers on AI pro­jects rece­ive 3.0%-3.2% higher wages than wor­kers on non-AI projects”.